The purpose of this work is to ascertain the numbers of alleles, allele frequencies, and allele frequency differences among American Indian tribes. The genetic systems being typed are the same as those being used currently in our genetic linkage analyses. Approximately 30 individuals from each of 20 tribes collected primarily at the Albuquerque Indian Hospital between the years 1992 and 1994 are being analyzed. Various tests for allele frequency differences between tribal groupings based on cultural and linguistic affinities are being performed. In order to more fully quantify isolate structure, and exploit such populations for linkage analyses, we have developed a maximum likelihood method to characterize populations by their levels of gene identity. We have applied this method to microsatellite typings for three American Indian and three European populations. Low gene identity was observed in Europeans, approximately 28%. By contrast, gene identity was higher in all American Indian populations, about 39%. In the past year, we have studied patterns of gene diversity in 7 samples from populations in the Southwest and Alaska. These populations all speak closely related Athabascan languages, in spite of their dispersed geographic locations. When compared to non-Athabascan speaking neighbors, we do not find a strong tendency for these populations to comprise a unified gene pool. Rather, geographic proximity is the best predictor of the genetic relationship. This is true regardless of the culture tradition shared or the language spoken. This information is important to genetic linkage and disease association studies on American Indians because failure to account for population diversity can result in false evidence for linkage, and allelic heterogeneity among groups can create spurious associations with disease.